Nomenclature

[trvrb]

The authors need to define their nomenclature regarding clade / lineage / type / strain. For instance, they say: "We find that 1 unit of antigenic drift within a clade increases incidence by 1.26 standard deviations, while 1 unit of drift in a clade of the same type decreases incidence by 0.61 standard deviations and 1 unit of drift in clades of differing type decreases incidence by 0.32 standard deviations. The term for across lineage Bt is small in magnitude." I had to read this many times before concluding that type refers to A or B, clade to refers to any of the four classifications. I am still not sure if lineage has a specific meaning or is simply a redundant word for clade.

[trvrb]

I know there were previous comments on the terminology here being confusing. The most difficult piece is not having a single standard term to refer to each of the 4 groups of viruses we're dealing with: A/H3N2, A/H1N1, B/Vic and B/Yam. H3 and H1 are referred to as "subtypes" and Vic and Yam are referred to as "lineages". Lineage is already problematic because it can refer to a number of different phylogenetic scales; any single branch on the tree is a lineage. It would be good to be more up front about nomenclature early in the MS. How about going with the following:

• lineage – A/H3N2, A/H1N1, B/Vic and B/Yam are all distinct influenza lineages; can use terminology like "within-lineage drift"
• clade – do not use
• subtype – only used to refer to A/H3N2 or A/H1N1
• type – only used to distinguish A and B
• branch – used when describing individual chains of infection, do not use lineage, instead use branch

[plemey]

That seems like a logical nomenclature proposal to me. With this in mind, I would recommend changes (in bold, did not intend to shout) along this line in the introduction paragraph to make this more explicit

Before 2009, there were four major influenza VARIANTS [to avoid the use of clades and lineages here] circulating within the human population: the H3N2 and H1N1 subtypes of influenza TYPE A, and the Victoria and Yamagata lineages of influenza TYPE B. In the case of influenza A, subtypes A/H3N2 and A/H1N1 refer to the genes, hemagglutinin (H or HA) and neuraminidase (N or NA), that are primarily responsible for the antigenic character of a strain. In the case of influenza B, THE TERM 'LINEAGE' IS USED [you could even add 'instead of subtypes] TO REFER TO Victoria (B/Vic) and Yamagata (B/Yam) AS antigenically distinct VARIANTS THAT diverged from a COMMON ANCESTOR prior to 1980 \cite{Rota90}.

And then remain consistent throughout the manuscript

[trvrb]

I agree that this cleans up the paragraph, but I'm concerned about using variants here and then using lineages later. I'd like to chose something and stick with it. Possibilities include:

• lineage – within-lineage drift, across influenza lineages
• variant – within-variant drift, across influenza variants
• virus – virus-specific drift, across influenza viruses

Thinking about this more now, I might be inclined to go with virus actually. John and Alan made some corrections to the text in this direction.

[trvrb]

We've cleaned up our language regarding clades and lineages in the text. We consider A/H3N2, A/H1N1, B/Vic and B/Yam to each represent separate lineages of influenza. We no longer refer to separate clades.